Changes in extreme climate events could have notable impacts on human mortality, regional economies, and natural ecosystems and therefore such events are of great public concern and interest. However, to understand and predict regional scale or even local scale climate extremes is a great challenge not only because they are rare but also because they are sometime, if not often, non-Gaussian distribution.

The Gaussian assumption has been widely adopted in many traditional statistical methods (e.g. linear regression) to estimate linear trends, diagnose physical mechanisms, or construct statistical prediction/downscaling models. Therefore, for non-Gaussian climate extremes, alternative approaches are needed. 

Dr. QIAN Cheng from Institute of Atmospheric Physics, Chinese Academy of Sciences and his Hong Kong and Macao collaborators proposed two approaches for statistical prediction/downscaling of non-Gaussian climate extremes under the framework of a Macao Meteorological and Geophysical Bureau project initiated by the World Meteorological Organization.

One approach used a multiple linear regression model after transforming the non-Gaussian predictant to a quasi-Gaussian variable and used Pearson’s correlation test/linear regression to identify potential predictors; and the other approach used a generalized linear regression model when the transformation was difficult and used a nonparametric Spearman’s correlation test to identify potential predictors.

These two approaches were illustrated by analyzing hot extremes in Macao, including hot days and hot nights, both of which were found to have a non-Gaussian distribution. Based on these two approaches, the physical processes responsible for the interannual and interdecadal variability of these two hot extremes were explored prior to the construction of a physically based statistical prediction model. These models can then be used to do future projections based on Representative Concentration Pathway (RCP) scenario datasets of numerical climate models and help the climate risk management and human adaptation. 

Schematic diagram of two proposed approaches for statistical prediction/downscaling of non-Gaussian climate extremes (Image by IAP)

This study has been published on the latest issue of Journal of Climate. 

Citation: Qian, C., W. Zhou, S. K. Fong, and K. C. Leong, 2015: Two approaches for statistical prediction of non-Gaussian climate extremes: a case study of Macao hot extremes during 1912−2012. J. Climate, 28(2), 623−636.